Association between Neuropsychological Performance and CSF Profile in Subjective Cognitive Decline: Towards the Diagnosis of Preclinical AD.

Background: In the perspective of novel treatments with disease-modifying drugs, a timely diagnosis of Alzheimer's' disease (AD) at preclinical phase represents a major issue. To this purpose, in clinical setting, there is the need to detect the earliest cognitive symptoms not yet fulfilling Mild Cognitive Impairment criteria, in order to proceed to biomarker assessment for diagnostic definition. In terms of cognitive performance, Subjective Cognitive Decline (SCD) is still a controversial entity, due to the difficulty of reliably measuring subtle deficits.

Objective: To evaluate the possibility to predict the presence of AD-like CSF pattern in SCD individuals, according to their neuropsychological performance assessed by means of both traditional and computerized measures.

Design: Retrospective study.

Setting: Clinical setting (Centre for Memory Disturbances, Section of Neurology, University Hospital of Perugia, Italy).

Participants: 74 consecutive SCD subjects who underwent an in-depth (paper-pencil and computerized) neuropsychological assessment and CSF analysis for AD biomarkers (Aβ42/Aβ40 ratio, phospho-tau, total tau).

Measurements: Neuropsychological assessment was composed of traditional tests assessing five cognitive domains (verbal memory, attention, executive functions, language, visuo-spatial abilities) and computerized tasks from CAmbridge Neuropsychological Test Automated Battery (CANTAB) (Pattern Recognition Memory, Paired Associates Learning and Spatial Working Memory). According to their performance at traditional tests, SCD individuals were categorized into cognitively normal (CN) and subtle impaired (SI); with respect to CANTAB, they were defined as CANTAB- in presence of normal performance, and CANTAB+ in presence of at least one pathological score. The subgroup with completely normal performance was defined as CN/CANTAB-, and the subgroup with impairment in both measures as SI/CANTAB+. Differences in prevalence of A/T/N profile according to cognitive profiles were assessed by Fisher's exact text for count data.

Results: None of CN/CANTAB- subjects showed A+/T+ status. SI/CANTAB+ subjects showed a significantly high prevalence of A+/T+ profile (14/35, 40%, p=0.03 vs CN/CANTAB-).

Conclusion: The neuropsychological profile may be of help in identifying SCD subjects requiring biomarker assessment. If confirmed in larger cohorts, the combination of traditional and computerized tests (namely, CANTAB) might represent a feasible strategy in clinical setting for carrying out biomarker assessment in individuals before the MCI stage. Detection of AD in these subjects would give them the highest chances to halt disease progression by means of disease modifying treatments.